DMS Colloquium: Dr. Guanqun Cao

Speaker: Dr. Guanqun Cao (Auburn)

Title: Statistical Learning for Next-Generation Complex Data Analysis

Abstract:  In the era of big data, advancements of modern technology have enabled the collection of sophisticated and ultra-high dimensional data sets, such as spatial data, 2D/3D images, and other objects. Most existing approaches, such as standard nonparametric smoothing, empirical likelihood, and quadratic discriminant analysis, perform well only under relatively simple data structure framework and strong assumptions. In this talk, I overview my recent research on the statistical learning for complex data analysis from three fundamental areas: regression, inference, and classification. In particular, newly developed varying coefficient Geo models, deep learning based multi-dimensional functional data regression, and classification methods will be presented.